Despite loud and frequent proclamations that cloud technologies are a new reality, that business must take place in the cloud, many large companies with legacy systems, some of which date back decades, do not say goodbye to these systems, on the contrary. They are using them and will continue to use them as part of a hybrid cloud strategy.
The reasons for using the legacy system in the digital age are often more than legitimate. For example, companies may have complex levels of integration into business systems that cannot be switched to the cloud with the lift and shift method. At the same time, the application of new enterprise systems is increasingly taking place in the cloud, creating so-called information silos: fragmented data within the legacy and the new, cloud environment. Additionally, analytics and BI teams have to deal with the challenges of integrating and analysing data from two or more data platforms. Numerous questions arise in this regard.
Hybrid data strategy in response to challenges
How to remove information barriers? How to fit different solutions and analytical platforms? Where to find support for the transformation of data strategy that affects business development, experience of employees but also end users?
Discussing hybrid data platforms, panellists Milan Bukorović, Chief Technology Officer at Nelt, Ivan Karlović, Director of Analytics and Data at Norwegian Airlines and Dimitar Dilov, Chief Risk Officer at Raiffeisen Bank shared their experiences gained through numerous enterprise transformation projects. The panel was moderated by Aleksandar Nedeljković, Chief Business Officer in Mainstream.
For Nelt, a leader in distribution and logistics, a data management strategy has proven to be a way to create added business value from a pool of information.
“Nelt has defined the need to conduct a digital maturity assessment, which we conducted in collaboration with Mainstream and Microsoft experts. One of the conclusions was that our company deals with a large amount of data from which it does not create business value, at least not completely,” explained Milan Bukorović.
According to him, NELT saw that the transformation of data, i.e the manner of data management and analytics, is the backbone of a broader transformation – the way of operation. But the road from achieving to setting goals is a long one and there are many factors to look out for such as knowing the data the company generates, form (whether structured or unstructured data), decisions about where to store and for what to use data.
Instead of combining data into one system, a hybrid data strategy allows companies to store data in structures that best suit their purpose. Ivan Karlović emphasized that the possibility of choosing data management today is greater than ever.
“Everything used to be kept on-premise. Today, there are great combinations of our own and cloud technologies, which Norwegian Airlines has recognized as strategically important, “said Karlović, illustrating with his own example that bad times are a good opportunity to consider business analytics.
“I joined Norwegian Airlines 17 days after the first Norwegian lockdown. Even then, in the unknown, our CEO for analysts said it was something we had to do. We need to build the foundations, choose the right platforms and technologies to create value later.”
How to choose the right technologies and partners for data transformation?
“If a tool cannot be easily integrated into our business, it is not a good tool for us,” said Dimitar Dilov. Speaking about the specifics of the banking sector he comes from, Dilov emphasized the importance of compliance with data-related regulations, such as the General Data Protection Regulation (GDPR). As he said, the requirements are very strict, while internal analyses can create compromises.
As it strives to meet the requirements of “two different worlds”, Raiffeisen Bank uses a wide range of analytical tools, as well as two cloud platforms: Amazon Web Services and Microsoft Azure. On the other hand, Norwegian Airlines is striving to establish a single cloud solution. For this company, the choice of cloud partners was based on the ecosystem of partners – providers of analytical solutions.
“The fact that we use the same cloud as our provider makes us more agile,” concluded Ivan Karlović, adding that platform support for the use of analytics is crucial.
Data platform architecture: ETL or ELT?
When it comes to data platform architecture, companies can choose between at least two paths – ETL and ELT. The main one between these approaches is where the data is transformed. In the first model (Extract, transform, load) the transformation occurs before the data is loaded into the Data Warehouse, while in the second (Extract, load, transform) the data is entered without any formatting and transformed in the Data Warehouse itself. How to choose an adequate model and what was the experience of a system like Nelt, which generates up to 10,000 transactions and invoices per day?
“ETL or ELT? It is individual and depends on the current environment, digital maturity and other factors. We had long discussions about that before we concluded that ELT suits our model,” underlined Milan Bukorović.
Perhaps the biggest advantage of the ELT model is the loading speed as well as the availability of raw data that has not been transformed before loading into the Data Warehouse. The final choice, however, depends on various factors such as the existing network infrastructure, budget and the extent to which the company already uses cloud and big data technologies.
Organization for optimal data strategy
The business community is aware that the most important thing for the use of analytical technology platforms are people. There is no doubt that the data scientist is a rising profession, but even users who do not specialize in the field of data need to improve their knowledge.
“We see an increasing demand for people who know new technologies such as the cloud and data processing tools. In the domain of risk, we started to motivate our people to go to trainings, get certificates. We want to internally increase the awareness of our people in the bank about data and to encourage them to develop data skills,” explained Dimitar Dilov.
The same line of “awareness raising” is applicable to company management, which sometimes remains in the dark about the value of business analytics. In that domain, as Ivan Karlović pointed out, small victories on the road to data transformation should be presented to the management. The last, but not the least, item is communication.
“Everything we do, we do for business, not technology. This is where we come to the challenge of change management – people are used to working in a certain way, they are used to certain tools,” said Milan Bukorović, adding that clear communication is necessary to introduce any changes in the work environment.
One of the conclusions of the panellists is that the organization of analytical projects is simpler today thanks to cloud technologies. Companies no longer have to engage in complex implementations of tools that they may not even use, but can try them out and dynamically change plans on the cloud. Cloud platforms certainly support agile development, reduce capital investment, and are one of the important elements of a hybrid data strategy. The only question is whether the company has the technological and industrial expertise to get the most out of the cloud and put it at the service of its data transformation. If you need support in this domain, our cloud experts are here to advise you.